3267RESEARCH ARTICLE
INTRODUCTIONNeural crest (NC) is a transient embryonic population, unique to
vertebrates, which arises at the border of the neural plate and gives
rise to a huge variety of cell types, such as neurons and glia in the
peripheral nervous system, connective tissues of the craniofacial
structures and pigment cells of the skin (Le Douarin and Kalcheim,
1999). This cell population is induced at the gastrula stage by signals
produced by the mesoderm, neural plate and epidermis (Raven and
Kloos, 1945; Selleck and Bronner-Fraser, 1996; Mancilla and
Mayor, 1996; Marchant et al., 1998; Bonstein et al., 1998; Monsoro-
Burq et al., 2003). Studies in chick, zebrafish and Xenopus embryos
have identified many NC-inducing signals, including BMPs, Wnts,
FGF, Notch and RA (reviewed by Basch et al., 2004; Steventon et
al., 2005). From all these molecules the most compelling evidence
exists for Wnt and BMP signals. Work in Xenopus and zebrafish
shows that inhibition of BMP signalling combined with Wnt
activation is required for NC induction (Saint-Jeannet et al., 1997;
La Bonne and Bronner-Fraser, 1998), and in chick Wnt signalling
has clearly been implicated in NC specification (Garcia-Castro et
al., 2002). However, the mechanism by which BMP and Wnt
signalling interact to induce NC is not clear. It has been proposed
that Wnts act as posteriorizing signals during this process, but this
remains controversial (Villanueva et al., 2002; Wu et al., 2005).
The anteroposterior axis of the neural plate is specified in a two-
step process (Nieuwkoop, 1952) (reviewed by Gamse and Sive,
2000). First, all neural tissue is induced as anterior neural plate; this
effect is driven by BMP antagonists. In the second step, called
transformation or posteriorization, the anterior neural plate is
posteriorized by Wnt, RA and FGF coming from the caudal end of
the embryo. Interestingly, the same combination of signals induces
NC (Villanueva et al., 2002). Moreover, only the posterior neural
fold develops into neural crest, whereas the anterior neural fold gives
rise to placodal tissue. In analogy with neural tissue, it has been
proposed that initially all the neural fold is induced with an anterior
(preplacodal) character dependent on attenuation of BMP signals
(Brugmann et al., 2004; Litsiou et al., 2005; Ahrens and Schlosser,
2005). Subsequent posteriorization by Wnt, FGF and RA signals
permits the development of NC tissue from the medioposterior
neural-fold region (Papalopulu and Kintner, 1996; Villanueva et al.,
2002). The anterior neural-fold area does not express NC markers,
but can be transformed into NC by the activation of Wnt, FGF or RA
(Villanueva et al., 2002). The role of posteriorization during neural
crest induction has been recently challenged. It has been shown that
activation of Wnt/β-catenin signalling leads to NC induction without
affecting the anteroposterior axis of the neural plate, suggesting that
NC induction is independent of posteriorization (Wu et al., 2005).
In order to explore further a possible role of posteriorization in
NC induction, we have analysed the function of the homeobox gene
Gbx2. Gbx2 has been implicated in mediating caudalization of the
neural plate by Wnt, FGF and RA, and in establishing the midbrain-
hindbrain boundary (MHB) (Simeone, 2000).
Here we show that Gbx2 is expressed in a broad ectodermal
domain, including the area from which NC arises. Loss-of-function
experiments show for the first time that Gbx2 is required for NC
induction, whereas gain-of-function experiments indicate that Gbx2induces NC genes. Our results indicate that Gbx2 is essential for the
anteroposterior division of the neural folds: expression of Gbx2
promotes posterior neural fold fate (NC) while inhibiting anterior
neural fold fate (preplacodal). Surprisingly we found that, although
controlled by the same signals, the anteroposterior specification of
the neural folds is independent from the anteroposterior specification
of the neural plate; reconciling previous discrepant results
(Villanueva et al., 2002; Wu et al., 2005; Carmona-Fontaine et al.,
2007). We show that the NC specification by Gbx2 depends on its
interaction with the neural fold gene Zic1. Zic1 alone drives the
expression of preplacodal genes, whereas a combination of Zic1 and
The posteriorizing gene Gbx2 is a direct target of Wntsignalling and the earliest factor in neural crest inductionBo Li*, Sei Kuriyama*, Mauricio Moreno and Roberto Mayor†
Wnt signalling is required for neural crest (NC) induction; however, the direct targets of the Wnt pathway during NC inductionremain unknown. We show here that the homeobox gene Gbx2 is essential in this process and is directly activated by Wnt/β-cateninsignalling. By ChIP and transgenesis analysis we show that the Gbx2 regulatory elements that drive expression in the NC responddirectly to Wnt/β-catenin signalling. Gbx2 has previously been implicated in posteriorization of the neural plate. Here we unveil anew role for this gene in neural fold patterning. Loss-of-function experiments using antisense morpholinos against Gbx2 inhibit NCand expand the preplacodal domain, whereas Gbx2 overexpression leads to transformation of the preplacodal domain into NCcells. We show that the NC specifier activity of Gbx2 is dependent on the interaction with Zic1 and the inhibition of preplacodalgenes such as Six1. In addition, we demonstrate that Gbx2 is upstream of the neural fold specifiers Pax3 and Msx1. Our results placeGbx2 as the earliest factor in the NC genetic cascade being directly regulated by the inductive molecules, and support the notionthat posteriorization of the neural folds is an essential step in NC specification. We propose a new genetic cascade that operates inthe distinction between anterior placodal and NC territories.
KEY WORDS: Neural crest induction, Posteriorization, Gbx2, Wnt, Pax3, Msx1, Zic1, Genetic cascade, Xenopus
Development 136, 3267-3278 (2009) doi:10.1242/dev.036954
Department of Cell and Developmental Biology, University College London,Gower Street, London WC1E 6BT, UK.
*These authors contributed equally to this work†Author for correspondence ([email protected])
Accepted 22 July 2009 DEVELO
PMENT
3268
Gbx2 inhibits this fate and induces NC genes. Furthermore, we show
that Gbx2 is an immediate direct target of the Wnt signalling, and
that β-catenin/TCF3 binds to the Gbx2 promoter. Importantly Gbx2exerts its action upstream of the earliest expressed NC genes Pax3and Msx1. Our results place Gbx2 on the top of the NC genetic
cascade, being the earliest factor that is directly regulated by the NC
inductive signals.
MATERIALS AND METHODSXenopus embryos, micromanipulation, whole-mount in situhybridization and cartilage stainingXenopus embryos were obtained as described previously (Gómez-Skarmeta
et al., 1998) and staged according to Nieuwkoop and Faber (Nieuwkoop and
Faber, 1967). Animal caps were performed as described by Mancilla and
Mayor (Mancilla and Mayor, 1996). For in situ hybridization, antisense
digoxigenin or fluorescein-labelled RNA probes were used. Specimens were
prepared, hybridized and stained using the method of Harland (Harland,
1991), and NBT/BCIP or BCIP alone were used as substrates for the alkaline
phosphatase. The genes analysed were Snail2 [formerly Slug (Mayor et al.,
1995)]; FoxD3 (Sasai et al., 2001); Gbx2 (von Bubnoff et al., 1996); Otx2(Blitz and Cho, 1995); Sox2 (Kishi et al., 2000); Six1 (Ghanbari et al., 2001);
Cpl1 (Richter et al., 1998); En2 (Hemmati-Brivanlou et al., 1991); Krox20(Bradley et al., 1993); Pax3 (Bang et al., 1997); Msx1 (Suzuki et al., 1997)
and cytokeratin Xk81A (Jonas et al., 1985); tbx2 (Takabatake et al., 2002).
For cartilage staining, embryos were fixed at stage 45-47 and processed as
previously described (Tribulo et al., 2004).
RNA synthesis in vitro and microinjection of mRNAsAll plasmids were linearized and RNA transcribed as described by Harland
and Weintraub (Harland and Weintraub, 1985), using SP6 or T7 RNA
polymerases, and the GTP cap analogue (New England Biolabs). After
DNAse treatment, RNA was purified (BD Biosciences) and resuspended in
DEPC-water for injection and lineage tracing. RNA was co-injected with
FLDx (Molecular Probes) using 8-12 nl needles into 8- or 32-cell embryos
as described by Aybar et al. (Aybar et al., 2003). The use of FDX as a lineage
marker allows the identification of the injected side of the embryo but not
the injected cells, as usually only the strongest fluorescent is visible and
many times the precipitate from the in situ hybridization quenches the
fluorescence. The constructs used were: DD1 (Sokol, 1996); Wnt8 (Sokol,
1996); β-catenin-GR (Domingos et al., 2001); Gbx2 (Glavic et al., 2002);
Gbx2EnRGR (Glavic et al., 2002); tBR [truncated BMP receptor (Graff et
al., 1994)]; Pax3 (Sato et al., 2005), HD-Msx1 (Tribulo et al., 2003); Msx1
(Tribulo et al., 2003); Zic1 (Nakata et al., 1998). Treatment with
dexamethasone was performed as described previously (Tribulo et al.,
2003).
Gbx2 morpholinosA translation-blocking morpholino (ATG MO) and a splicing MO (Spl MO)
against Gbx2 were designed, and both of them produced the same neural
crest phenotype. The sequence of the translation-blocking MO (ATG MO)
is: 5�-GCTGAAAGGCTGCACTCATATAAGC-3�. A seven-mismatch
Gbx2 mutant (7mutGbx2) was generated to perform the rescue of the ATG
MO. To amplify the 7mutGbx2, the following primers were used: forward
primer: 5�-GAATTCACC ATGtcgGCtGCtTTcCAa CCCCCTCTCATGAT-
3� and reverse primer: 5�-TTGACTCGAGTCAAGGTCTT GCTT -
GCTCCAGC-3�. The italic letters show the morpholino target sequences.
The lower case shows mismatching nucleotides inserted by PCR without
changing amino acid sequences. The amplified Gbx2 open reading frame
(ORF) cDNA was subcloned into pCS2+ expression vector. To design the
splicing MO we cloned the first intron of the Xenopus laevis Gbx2 gene. We
predicted the structure of X. laevis Gbx2 allele from that of X. tropicalisGbx2. The primers for exon1 (5�-GAGAAGAAGGAA ACAA -
GACCTACAT-3�) and exon2 (5�-AGTTTGGCAGGAGATATTGTCATCT-
3�) were designed to amplify 200 bp in the ORF of cDNA, and 1.7 kb in
genomic DNA. The isolated genomic fragment was subcloned in
pBluescript II vector, and the sequences of exon-intron boundary were read
from several different clones. The sequence of the splicing MO is: 5�-GTGATGGTTGCTACACTTACCTAGA-3� (Gene Tools). To check the
efficiency of splicing MO, we performed RT-PCR with exon1-exon2
primers as described above. The shifted RNA was not amplified by RT-PCR
using Go-Taq (Promega) (data not shown).
Chromatin immunoprecipitation and promoter analysisChromatin immunoprecipitation (ChIP) was based on the method described
previously (Stewart et al., 2006). Briefly, 100 embryos from each stage
(stages 11 and 14) were homogenized in 1 ml of nuclei extraction buffer, and
fixed in 1% formaldehyde/nuclei extraction buffer A [0.5% TritonX-100, 10
mM Tris-HCl (pH 7.5), 3 mM calcium chloride, 0.25 M sucrose, 1 mM
DTT, 0.2 mM PMSF, 1 tablet of complete mini (Roche applied science)] at
room temperature for 15 minutes. The fixation was quenched by 0.125 M
glycine for 5 minutes. The lysate was filtered with 100 μm cell strainer,
centrifuged at 2000 g, resuspended in 1 ml of nuclei extraction buffer C [10
mM Tris-HCl (pH 7.5), 3 mM calcium chloride, 2.2 M sucrose, 1 mM DTT,
0.2 mM PMSF], layered on 9 ml of nuclei extraction buffer B [0.5%
TritonX-100, 10 mM Tris-HCl (pH 7.5), 3 mM calcium chloride, 2.2 M
sucrose, 1 mM DTT, 0.2 mM PMSF], and the nuclei were extracted by
ultracentrifugation (40,000 g, 4°C, 3 hours). The nuclei were re-suspended
in 3 ml SDS lysis buffer. Chromatin-protein complexes were sonicated in 15
ml Falcon tubes, and 100 μl aliquots were flash frozen in liquid nitrogen.
The DNA concentration in one of the aliquots was determined after
proteinase K treatment. ChIP was done with Protein G-dynabeads
(Invitrogen). Protein G-dynabeads were blocked with 200 μg/ml salmon
testis DNA (SIGMA) and 0.5 mg/ml BSA in ChIP buffer at 4°C overnight.
All the samples were diluted by adding ChIP buffer and pre-cleared with
Protein G-dynabeads. After pre-clear 20 μl of chromatin was taken as an
input control. Antiserum against β-catenin (final 1:1000, anti-rabbit
polyclonal, Calbiochem ab6302) was used to precipitate DNA-Tcf/Lef/β-
catenin complex, and pan-cadherin type I antiserum was used as an IgG
negative control (final 1:1000, anti-rabbit polyclonal, Sigma C3678).
Antibodies were incubated with Protein G-dynabeads at room temperature
for 30 minutes, and washed once. Then, 1 ml of 3-5 ng/μl (estimated) of pre-
cleared chromatin was added to 40 μl slurry of antibody-Protein G beads
complex. The eluted samples were treated with RNase A (RPA grade,
Promega) at 37°C for 15 minutes and proteinase K (Sigma) and incubated
at 65°C overnight. The DNA was purified using the MinElute Reaction
Cleanup Kit (QIAGEN), and the target sequences were amplified by Go-Taq
Flexi (Promega). PCR cycles; 94°C for 30 seconds, 57°C for 30 seconds,
72°C for 30 seconds, for 32-34 cycles. The primers used for ChIP are as
follows: Gbx2-ChIP-U: 5�-GAATCCCAACCACAGAAGGA-3�, Gbx2-
ChIP-D: 5�-GTAGGCACCAGAGCCACAGT-3� (–346 to –75, 272 bp/32
cyc); Snail2A-3kb U: 5�-TGTTTCCATCCCAACACCTG-3�, Snail2A-3kb
D: 5�-CTTTCACAGGCTGAGGCATT-3� [–3005 to –2802, 204 bp/33 cyc,
GenBank AF368040.1 (Vallin et al., 2001)]; Xen-2 GS-3 U: 5�-GGCT -
CGAGGCTTCTCCCCAAGCCC-3�, Xen-2 GS-3 D: 5�-CCAAGCTT -
GTCGCCTCCAGACCCG-3� (33 cyc). This GS primer is used for
amplification of competitor fragments in gel-shift assay as described in
McGrew et al. (McGrew et al., 1999); alpha tubulin intron-U: 5�-TGAA -
ACAGGAGCAGGAAAGC-3�, alpha tubulin intron-D: 5�-GCTCT GG -
GTGGAATAACTGC-3� (34 cyc) (Dunican et al., 2008).
For promoter analysis the 5� upstream region of Gbx2 was cloned and
fused to GFP as a reporter gene. Different deletion constructs were injected
into Xenopus embryos, and transient transgenics were generated using the
Tol2 system (Kawakami, 2007).
RNA isolation from embryos and RT-PCR analysisTotal RNA was isolated from whole embryos or embryonic tissues after
microdissection and cDNA were synthesized as previously described (Aybar
et al., 2003). The primers used in this study were: Gbx2U 5�-AAACTGCCCACAAAGAGGAGGAC-3�, Gbx2D 5�-TGGTGTTGGC -
TCCGTATGGCAAA-3�; Six1 (Pandura and Moody, 2000); ODC (Heasman
et al., 2000); Pax3 (Hong and Saint-Jeannet., 2007); Nrp1 (Richter et al.,
1990); Snail2 (Aybar et al., 2003); FoxD3 (Sasai et al., 2001); Sox9 (Monsoro-
Burq et al., 2003); EpK (Jonas et al., 1989). PCR amplification with these
primers was performed over 23-27 cycles and the PCR products were analysed
on 1.5% agarose gels. As a control, PCR was performed with RNA that had
not been reverse-transcribed to check for DNA contamination.
RESEARCH ARTICLE Development 136 (19)
DEVELO
PMENT
RESULTSGbx2 is expressed in the prospective neural crestWe hypothesized that the NC is induced in posterior ectoderm by
Wnt signals (Fig. 1A) and that the posteriorizing activity of Wnt
is mediated by Gbx2 (Fig. 1C). An essential aspect of this
proposal is the demonstration that Gbx2 is expressed in the NC at
the time of its induction (Fig. 1B). The expression of Gbx2 was
compared with the expression of the NC marker Snail2. At the
late gastrula stage (stage 12), when NC induction takes place
(Mancilla and Mayor, 1996), Snail2 and Gbx2 were clearly co-
expressed by NC cells (Fig. 1D,E). This was further confirmed by
double in situ hybridization for Snail2 and Gbx2 (Fig. 1F,G).
Sections show that Gbx2 expression was restricted to the
ectoderm; it was not expressed in the mesoderm and it overlapped
with Snail2 expression (Fig. 1H-K). Gbx2 was expressed in a
wide domain of ectoderm, which included the prospective NC
domain (Fig. 1L-O) and was absent only from the anteriormost
region of the embryo (Fig. 1M-O). It should be noted that later in
development, at the neurula stages, the expression of Gbx2 was
inhibited in the midbrain posterior domain (Glavic et al., 2002);
however, at gastrula stages this area still expressed Gbx2 and
included the most anterior prospective NC cells (Fig. 1F,N)
(Glavic et al., 2002). We noticed at the neurula stage a gap
between the neural and epidermal expression of Gbx2 (Fig. 1Q,
arrow). Double in situ hybridization with Snail2 showed that this
gap in Gbx2 expression was adjacent to the NC (Fig. 1R-T),
corresponding presumably to the preplacodal region (PPR). In
summary, Gbx2 is expressed in the prospective NC territory at the
time of NC induction (Fig. 1B).
Gbx2 is essential for NC inductionIn order to analyze Gbx2 function in NC induction, we undertook
an MO loss-of-function approach. MOs were injected into one
animal blastomere of eight-cell stage embryos. In all injected
embryos shown in this work the injected side is shown to the right
side. FLDx was used to recognize the injected side. Although
injection of control MO did not affect the expression of Snail2(Fig. 2A,I), injection of splicing (Spl) – or translation (ATG)
blocking – MOs produced a strong inhibition of this NC marker
(Fig. 2B,C,I).
Injection of Gbx2 mRNA led to a modest but consistent
expansion of Snail2 expression (Fig. 2D); a similar expansion was
observed when mRNA coding for Gbx2EnR [Gbx2 homeodomain
fused to Engrailed repressor domain of Drosophila (Glavic et al.,
2002)] or Gbx2 mutated in the MO-binding site were injected (Fig.
2F,H). Injection of Gbx2EnR or the mutated Gbx2 mRNA were able
to rescue the inhibition of Snail2 expression produced by translation-
blocking MO (Fig. 2E,G,I), showing its specificity.
To measure the efficiency of the MO treatment we analyzed the
splicing of Gbx2 when embryos were injected with control MO or
Gbx2 splicing MO. A strong inhibition of Gbx2 splicing was
observed with the Spl MO (Fig. 2J). As the splicing and translational
MO produced the same effect on NC induction we used the
translational MO in most of the subsequent experiments (from now
on referred to as Gbx2 MO).
As Gbx2 is expressed in the neural plate and epidermis in addition
to the NC, we performed targeted injections of Gbx2 MOs at the 32-
cell stage into blastomeres fated to become neural plate, NC or
epidermis (Fig. 2M-P). Only Gbx2 depletion in the neural crest was
able to inhibit Snail2 (Fig. 2L,O), and no effect was observed when
injected into the neural plate (Fig. 2M,O) or epidermis adjacent to
the NC (Fig. 2N,O).
Gbx2 is a direct target of Wnt signalling duringneural crest inductionWe previously hypothesized that one of the first steps in NC
induction was posteriorization of anterior ectoderm by NC-inducing
signals (Aybar and Mayor, 2002). As Gbx2 has been involved in
posteriorization, we propose that it could be one of the earliest
factors activated in the NC induction cascade. Accordingly, we
asked whether Gbx2 is a direct target of NC-inducing/posteriorizing
signals.
Activation of Wnt signalling by injection of Wnt8 or β-catenin-
GR mRNA in the prospective NC region led to expansion of
endogenous Snail2 expression (Fig. 3A), whereas inhibition of Wnt
signalling by expression of a dominant negative of Dsh (DD1)
(Tada and Smith, 2000) led to NC inhibition (Fig. 3B) (DeCalisto
et al., 2005). A similar expansion or inhibition of Gbx2 was
observed after activation or inhibition of the Wnt pathway (Fig.
3C,D), suggesting that NC induction by Wnt may be Gbx2
dependent. We tested whether Gbx2 was a direct target of Wnt
signalling during neural crest induction by inhibiting protein
3269RESEARCH ARTICLEGbx2 in neural crest induction
Fig. 1. Gbx2 is expressed in posterior ectoderm that includes theprospective neural crest. (A-C) Hypothesis of neural crest (NC)induction by the posteriorizing activity of Gbx2. (D-S) In situhybridization at the indicated stages for the indicated genes.(D-G) Dorsal view, anterior to the top. (H-K) Transverse sections.(L-O) Lateral view, anterior to the left. (P-R) Dorsal view, anterior to thetop. Arrowhead, NC; arrow, gap in Gbx2 expression. (S) Detail of theneural fold region in a lateral view, anterior to the top, midline to theright. (T) Summary of Gbx2 and Snail2 expression at stage 16. Anteriorto the top, midline to the right. Different tones of purple denotedifferent levels of Gbx2 expression. Blue, NC.
DEVELO
PMENT
3270
synthesis in explants previously injected with a combination of tBRand β-catenin-GR mRNAs to induce NC. Injected animal caps were
treated with the protein synthesis inhibitor cyclohexamide (CHX)
at stage 11.5, and 30 minutes later β-catenin activity was triggered
by addition of dexamethasone (DEX). After 2 hours of culture,
mRNA was extracted and RT-PCR was performed for Gbx2, Sox9and Snail2.
NC induction by inhibition of BMP and activation of Wnt
signalling was confirmed by Sox9 and Snail2 expression after DEX
treatment (Fig. 3E, fourth lane). In the same situation, Gbx2expression was induced (Fig. 3E). However, treatment with CHX
completely blocked Sox9 expression (Fig. 3E, sixth lane), indicating
that protein synthesis is being inhibited and that it is required for
Sox9 expression. A strong reduction in Snail2 expression was
observed, suggesting that most of the Snail2 expression analysed at
this early stage is not directly regulated by Wnt signalling, as it can
be deduced from several reports that place Snail2 downstream of
Pax3, Msx1 and Zic1 (Kuo et al., 1998; Tribulo et al., 2003;
Monsoro-Burq et al., 2005; Sato et al., 2005; Hong and Saint-
Jeannet, 2007; Zhao et al., 2008) (and see Discussion). Interestingly,
no effect on Gbx2 expression was observed in the presence of CHX
(Fig. 3E, sixth lane). These results show that once Wnt signalling is
activated protein synthesis is not required for Gbx2 upregulation,
suggesting that Gbx2 transcripts are directly regulated by β-catenin.
To further investigate whether Gbx2 is directly regulated by Wnt
signalling, we analysed the presence of TCF/LEF binding sites in
the 5� region of the Gbx2 gene. We identified a region of 500 bp
upstream of Gbx2 that contains three (1-3) LEF/TCF consensuses
(Fig. 3A), with site 1 as a perfect TCF/LEF consensus (Eastman and
Grosschedl, 1999).
RESEARCH ARTICLE Development 136 (19)
Fig. 2. Gbx2 is required for NC induction. Embryos were injectedin animal blastomeres at the eight-cell stage with the indicated MO,and the expression of Snail2 was analysed between stages 12 and13. In all the images, in situ hybridizations are shown in dorsal viewwith anterior to the top and the inset corresponds to the overlay ofin situ hybridization and fluorescence to show the injected side tothe right. (A) Control MO (20 ng). (B) Gbx2 splicing MO (20 ng).(C) Gbx2 translational MO (16 ng). (D) Gbx2 mRNA (1 ng). (E) Gbx2translational MO (16 ng) and Gbx2EnR-GR (1 ng). Dexamethasonewas added at stage 10. (F) Gbx2EnR-GR (1 ng). Dexamethasone wasadded at stage 10. (G) Gbx2 translational MO (16 ng) and seven-mismatch (7mismatch) Gbx2 mRNA (1 ng). (H) 7mismatch Gbx2mRNA (1 ng). (I) Summary of rescue experiment showing percentageof embryos with Snail2 inhibition. ** P<0.001. (J) Efficiency ofsplicing MO. RT-PCR of embryos injected with 20 ng of control MOor 20 ng of Gbx2 splicing MO. Gbx2 and ODC were analysed. ODC,loading control. (K-N) Targeted injection of Gbx2 translational MO.A1, A3 or A4 blastomeres were injected with Gbx2 MO to targetneural plate, NC or epidermis, respectively. (K) NC injection.(L) Neural plate injection. (M) Epidermis injection. (N) Summary oftargeted injection, showing percentage of Snail2 inhibition afterinjecting in prospective NC, neural plate or epidermis. A minimum of30 embryos was analysed in each experiment. E, epidermis; NP,neural plate.
Fig. 3. Gbx2 is a direct target of Wnt signaling. (A,C) Embryoinjected in animal blastomeres of an eight-cell-stage embryo with 1 ngof Wnt8 mRNA. (B,D) Embryo injected with 1 ng of DD1 mRNA intoanimal blastomere of an eight-cell embryo. Snail2 (A,B) or Gbx2 (C,D)expression was analysed at stage 12.5. (E) RT-PCR of animal capsanalysing Gbx2, Snail2 and Sox9 expression. AC, animal cap; CHX,cyclohexime added 0.5 hours before DEX; ctBR, 1 ng of dominant-negative of BMP4 receptor; DEX, dexamethasone added at stage 11.5;ODC, loading control; WE, whole embryo.
DEVELO
PMENT
To analyse whether this 500 bp region of Gbx2 works as a
regulatory region, we fused it to GFP (Fig. 3B) and transient
transgenic frogs were generated. Fluorescence analysis shows a
clear expression of GFP in the NC region (Fig. 3C, dashed line), in
addition to other domains that are consistent with Gbx2 expression.
This observation suggests that the 500 bp contains NC regulatory
elements of Gbx2, and we will call it a Gbx2 enhancer. To test
whether the Gbx2 enhancer responds directly to Wnt signalling, we
induced NC in animal caps taken from transgenic embryos injected
with a combination of tBMPR and inducible β-catenin, as described
for Fig. 3E. The presence of GFP protein was analysed by western
blot and Gbx2 by RT-PCR (Fig. 4D). Activation of β-catenin by
dexamethasone treatment led to a strong increase in GFP protein and
Gbx2 mRNA, but only GFP protein was sensitive to the protein
synthesis inhibitor CHX (Fig. 4D, lane 3). This experiment suggests
that the Gbx2 enhancer is able to respond to Wnt signalling and that
this response is direct, as no protein synthesis is required.
In order to identify which of the three putative TCF/LEF binding
sites of the Gbx2 enhancer were functional, we performed specific
deletions for each of them (Fig. 4E) and tested their activity in
transgenic embryos (Fig. 4F). Control embryos (Fig. 4F, white bars)
or embryos injected with β-catenin (Fig. 4F, black bars) were co-
injected with each deletion construct, and GFP fluorescence was
analysed. We identified site 1 as essential for the β-catenin response
in the NC cells (Fig. 4E,F).
Then, to analyse whether β-catenin/TCF interacts physically
with site 1 of the Gbx2 enhancer, we conducted ChIP assays using
chromatin extracted from embryos at two different stages and an
antibody against β-catenin. We found that at the early gastrula
stage the antibody specifically precipitated site 1 of Gbx2, but did
not precipitate the Snail2 promoter or a tubulin intron used as
control (Fig. 4G). These findings indicate that β-catenin/TCF
does indeed associate physically with the putative Gbx2regulatory region at the early gastrula stage. We performed a
second ChIP at neurula stage (stage 14) and found that in addition
to Gbx2 precipitation, the Snail2 promoter was also precipitated
(Fig. 4H). Quantification of these results (Fig. 4I), including En2enhancer as positive and tubulin intron as negative control,
respectively, showed that β-catenin/TCF binds to the Gbx2 gene
during early NC induction, and later binds to the Snail2 promoter,
3271RESEARCH ARTICLEGbx2 in neural crest induction
Fig. 4. Analysis of Gbx2 regulatory region. (A) 5� region upstream of Gbx2. Boxes 1 to 3 indicate TCF/LEF consensus binding sites. Start codon isunderlined. (B) Fusion construct of Gbx2 putative enhancer (containing the three TCF/LEF consensus sites) and GFP as a reporter gene. (C) Transgenicembryos were generated and GFP fluorescence was visualized. A stage 12 embryo is shown. Dashed line, prospective NC. (D) Animal caps taken fromtransgenic embryos injected with tBMPR and β-catenin-GRto were treated with cyclohexamide and 0.5 hours later with dexamethasone. GFP proteinwas assayed by western blot and Gbx2 expression by RT-PCR. ODC, loading control. Note that CHX did not inhibit Gbx2 transcription but inhibitedGFP synthesis. (E) Deletion constructs used to develop transgenic embryos. Red X indicates deletion in the TCF/LEF binding site. (F) Percentage oftransgenic embryos showing GFP expression. Embryos were injected with β-catenin-GR, dexamethasone was added at stage 11 and GFP fluorescencewas analysed at stage 12. White bar, GFP fluorescence in absence of dexamethasone; black bar, GFP fluorescence in the presence of DEX. Note thatthe TCF/LEF binding site 1 is essential for Gbx2 enhancer activity. (G-I) ChIP assay. (G) ChIP assay on chromatin of stage 11 embryos. (H) ChIP assay onchromatin of stage 11 and 14 embryos. (I) Quantification of fold enrichment of ChIP at stages 11 (black bars) and 14 (gray bars) for Gbx2 and Snail2.En2 enhancer, positive control; Tubulin intron, negative control. βC, β-catenin antibody; CHX, cyclohexamide; DEX, dexamethasone; Gbx2, Gbx2enhancer; Ig, IgG pan-cadherin antibody; in, input; Snail2, Snail2 promoter; Xtub-Int, Tubulin intronl.
DEVELO
PMENT
3272
as has been shown by Vallin et al., 2001 (Vallin et al., 2001).
Taken together, these results demonstrate that Gbx2 is a direct
target of Wnt signalling.
Next we tested whether NC induction by Wnt signalling is Gbx2dependent. Injection of β-catenin-GR (DEX stage 11) led to an
expansion of the NC markers Snail2, Pax3 and Msx1 (Fig. 5A-C).This effect was reverted in the presence of Gbx2 MO (Fig. 5D-F).
Moreover, inhibition of the NC markers Snail2, Pax3 and Msx1 (Fig.
5G-I) by Wnt inhibition (Dsh dominant-negative DD1) was rescued
by co-injection of Gbx2 mRNA (Fig. 5J-L). Taken together, these
results show that NC induction by Wnt signalling is Gbx2 dependent
and makes Gbx2 the earliest target of this pathway during this
process.
Gbx2 is upstream in the NC genetic cascadeIn the genetic cascade controlling NC development Pax3 and Msx1have been proposed as the first factors activated by secreted
inducing signals (Mayor and Aybar, 2001; Meulemans and Bronner-
Fraser, 2004; Steventon et al., 2005). As Gbx2 is a direct target of
the NC inducer Wnt, we asked whether Pax3 and Msx1 expression
are regulated by Gbx2. First, we observed that Gbx2 was expressed
earlier than Pax3/Msx1 in the prospective NC area (Fig. 6A-F).
Second, loss of Gbx2 led to inhibition of Msx1 and Pax3 expression
(Fig. 6G-J), but the effect was stronger when the embryos were
analysed at the gastrula stage (stage 11; Fig. 6G,H), than at the
neurula stage (stage 14, Fig. 6I,J). We also observed that
overexpression of Gbx2 mRNA produced an expansion of NC
markers (not shown). In order to determine when Gbx2 was required
for NC development, an inducible Gbx2 construct was used, in
which Gbx2 was fused to the repressor domain of engrailed and to
the glucocorticoid receptor elements, which can be exogenously
activated (Glavic et al., 2002). Activation of Gbx2Enr-GR affected
the expression of the NC marker Snail2 only if it was activated
before the end of gastrulation (stage 12), with no effect after stage
15 (Fig. 6K). Taken together, these results indicate that Gbx2 is
required for the early phase of NC specification and not for the late
maintenance after stage 15 (Steventon et al., 2009).
Next we performed a series of epistasis experiments aimed to
confirm that Gbx2 was upstream of Pax3 and Msx1. Inhibition of
Snail2 and FoxD3 by Gbx2 MO (Fig. 7A,D) could be almost
completely rescued by co-injection of Pax3 or Msx1 mRNA (Fig.
7B,C,E,F). However, inhibition of NC induction by loss of function
of Pax3 (Fig. 7G) or Msx1 (Fig. 7J) could not be rescued by co-
injection of Gbx2 mRNA (Fig. 7H,I,K,L). Taken together, these
results indicate that Gbx2 is upstream of Pax3 and Msx1 in the NC
genetic cascade.
Gbx2 works as a posteriorizing factor of theneural foldPrevious evidence demonstrates that Gbx2 acts as posteriorizing
factor in the neural plate (Wurst and Bally-Cuif, 2001). Therefore,
we investigated whether Gbx2 also interferes with anteroposterior
patterning of the neural folds. As shown above, NC markers were
inhibited by Gbx2 MOs (Fig. 8A,C) and slightly expanded by Gbx2
overexpression (Fig. 8B,D). Interestingly, these effects were
accompanied by changes in the expression of the preplacodal marker
Six1. Six1 is normally expressed in the PPR of the anterior neural
fold. Upon Gbx2 MO injection, Six1 expression was expanded
posteriorly (Fig. 8E), and shifted anteriorly if Gbx2 was
overexpressed (Fig. 8F). Gbx2 repression of Six1 was further
confirmed in explant experiments. An animal cap explant expressed
Six1 when BMP signalling was inhibited [injection of a dominant-
negative form of the type II BMP receptor (tBR)] (Fig. 8G)
(Brugmann et al., 2004). However, this effect was reverted when
animal caps were co-injected with Gbx2 mRNA (Fig. 8G, fourth
lane). These experiments suggest that Gbx2 is a repressor of the
PPR.
These results indicate that Gbx2 is a posteriorizing factor for the
neural fold. We then asked whether the anteroposterior axis of the
neural plate is also affected in these conditions. Surprisingly, we found
that the anteroposterior patterning of the neural plate was normal.
Regionally restricted neural plate markers such as Cpl1, En2 and Otx2were expressed normally after similar Gbx2 MO or Gbx2 mRNA
injections (Fig. 8H-M; see Fig. S1A,C in the supplementary material).
Only much higher levels of Gbx2 MO than those used here affect
neural plate patterning (see Fig. S1B in the supplementary material).
In addition, Gbx2 knockdown did not change the expression of the
pan-neural plate marker Sox2, or the epidermal marker Keratin (Fig.
8N,O). In conclusion, reduction of Gbx2 produces an enlargement of
RESEARCH ARTICLE Development 136 (19)
Fig. 5. NC induction by Wnt signalling is Gbx2 dependent.Embryos were injected in animal blastomeres at the eight-cell stage asindicated. The expression of Snail2, Pax3 and Msx1 was analysed atstage 12. (A-C) β-catenin-GR (1 ng) and induced at stage 10 with DEX.Between 75 and 86% of NC expansion; n=153 embryos. (D-F) β-catenin-GR (1 ng), induced at stage 10 with DEX and 16 ng of Gbx2MO. NC expansion was reduced to less than 2%; n=174. (G-I) 1 ng ofDsh dominant-negative DD1. Between 78 and 82% of inhibition of NCgenes; n=124. (J-L) Dsh dominant-negative DD1 (1 ng) and 1 ng ofGbx2 mRNA. NC inhibition was reduced to less than 2%; n=120.
DEVELO
PMENT
the PPR at the expense of NC cells (Fig. 8P), whereas Gbx2
overexpression has the opposite effect (Fig. 8Q). Moreover, the
posteriorizing effect of Gbx2 in the neural folds is independent of the
anteroposterior patterning of the neural plate.
These observations suggest that Gbx2 is involved in the
anteroposterior differences of the neural fold, as the absence of Gbx2is required for specification of anterior placodes and its presence is
needed for development of NC and probably posterior placodes that
are at the same anteroposterior level as the NC, such as the otic
placode. NC and placode derivatives were analysed after interfering
with Gbx2 activity. Gbx2 MO produced a strong inhibition of NC
derivatives, such as cartilage and melanocytes (see Fig. S2A,B,E in
the supplementary material), while that Gbx2 mRNA injection leads
to inhibition of anterior placode such as lenses without major effect
in more posterior placodes such as otic (see Fig. S2C,D in the
supplementary material).
Gbx2 interacts with Zic1 to induce NCWe have shown that Gbx2 is essential to induce NC in the
posterior neural folds. However, our data suggest that Gbx2 is not
sufficient for NC induction, as its overexpression only can expand
the NC in the neural fold region. For this reason, we hypothesized
that Gbx2 may interact with another factor in the neural folds to
induce NC. As there is evidence that attenuation of BMP
signalling is necessary for neural fold specification (Marchant et
al., 1998; Nguyen et al., 1998; Streit and Stern, 1999), we
3273RESEARCH ARTICLEGbx2 in neural crest induction
Fig. 6. Early requirement of Gbx2 for NC induction. (A-F) In situhybridization for Gbx2, Pax3 and Msx1 at the indicated stages. Notethat only Gbx2 is observed at stage 11. (G-J) Gbx2 MO was injected atthe eight-cell stage and the expression of Pax3 and Msx1 was analysedat the indicated stages. Asterisks indicate the injected side (visualized inthe inset). Note almost complete inhibition of Pax3 and Msx1 at stage11.5 (43% of total and 31% of partial inhibition; n=83), and partialinhibition at stage 14 (70% of partial inhibition; n=110). (K) Percentageof embryos with defects in Snail2 expression after activation ofGbxEnR-GR with dexamethasone at the indicated stages (s).
Fig. 7. Gbx2 is upstream of Pax3 and Msx1 in the NC geneticcascade. Embryos were injected as indicated and the expression of theindicated genes was analysed at stage 12. (A-F) Gbx2 MO alone (16 ng)(A,D) or co-injected with 1 ng of Pax3 mRNA (B,E) or 1 ng of Msx1mRNA (C,F). Eighty-one to 83% of NC inhibition by Gbx2 MO (n=178)was rescued to less than 1% of inhibition by co-injection with Pax3(n=78) or Msx1 (n=69) mRNA. (G-I) Pax3 MO alone (20 ng) (G) or co-injected with 1 ng of Gbx mRNA (H,I). Seventy-eight percent (n=90) ofNC inhibition by Pax3 MO was not rescued by Gbx2 MRNA (75-79%inhibition; n=189). (J-L) Msx1 dominant-negative HD-Msx1 alone (1 ng)(J) or co-injected with 1 ng of Gbx2 mRNA (K,L). Sixty-eight percent(n=89) of NC inhibition by HD-Msx1 alone was not rescued by Gbx2mRNA (73-77% of inhibition; n=124).
DEVELO
PMENT
3274
reasoned that a factor induced by BMP attenuation could be
required together with Gbx2 for NC induction. Indeed, although
NC markers were not induced in animal caps by BMP inhibition
alone (Fig. 9A, third lane), inhibition of BMP in the presence of
Gbx2 led to the induction of NC markers (Fig. 9A, fifth lane),
similar to the effect obtained by BMP inhibition and Wnt
activation (Fig. 9A, fourth lane). The latter was completely
abolished when Gbx2 was inhibited by Gbx2 MO (Fig. 9A, sixth
lane). Taken together, our experiments indicate that attenuation of
BMP induces a factor that in turn interacts with Gbx2, which is
induced by Wnt signalling, to activate NC marker expression.
In order to identify the unknown factor, we tested several
candidate genes expressed in the early neural folds for their ability
to induce NC markers when co-expressed with Gbx2 in animal caps.
From these, only Zic1 was able to do so (Fig. 9B). In addition, we
observed that the induction of Six1 by Zic1 (Fig. 9B, fourth lane)
was completely inhibited by co-injection of Gbx2 (Fig. 9B, fifth to
seventh lanes). These results suggest that Zic1 is at least one of the
factors that interact with Gbx2 to induce NC in the posterior neural
fold and to inhibit preplacodal fates.
DISCUSSIONWe found that Gbx2 plays an essential role in NC induction and
propose a new genetic cascade that operates in the distinction
between NC and anterior PPR. Gbx2 resides at the top of the NC
genetic cascade, being directly activated by the NC inducer Wnt.
Several lines of evidence support this idea: (1) Gbx2 is expressed in
the prospective NC at the time of its induction; (2) inhibition of
Gbx2 expression leads to a complete loss of the NC region,
concomitant with an expansion of the PPR; (3) Gbx2 is a direct
downstream target of Wnt/β-catenin signalling during NC induction;
(4) Gbx2 is upstream of the earliest transcription factors, Pax3 and
Msx1, in the NC genetic cascade; and finally, (5) interaction between
Gbx2 and Zic1 is sufficient to induce NC and inhibit the PPR in
animal caps.
We propose the following model of NC induction (Fig. 9C-F).
Initially, a specific level of BMP activity induces Zic1 along the
entire neural plate border (Fig. 9C,F). The direct regulation of Zic1by BMP has been previously reported (Mizuseki et al., 1998; Rohr
et al., 1999; Tropepe et al., 2006; Hong and Saint-Jeannet, 2007).
Then, Wnt signalling induces directly the expression of Gbx2 in the
posterior region of the embryo (Fig. 9D,F). Zic1 by itself induces
Six1 and specifies the anterior placodal domain next to the anterior
neural folds (Fig. 9E,F), whereas Zic1 in combination with Gbx2induces NC (and posterior placode, such as otic) in the posterior
neural folds, where both genes are co-expressed (Fig. 9E,F). In
addition, Gbx2 inhibits Six1 expression and the PPR (Fig. 9F). In
summary, the presence of Gbx2 at the neural plate border defines the
region that becomes NC, and in its absence the region develops into
anterior placode territory.
RESEARCH ARTICLE Development 136 (19)
Fig. 8. Gbx2 is required for the posteriorization ofneural folds. Embryo injections and expression ofindicated genes were analysed at stage 12. (A,B) Snail2.(C,D) FoxD3. (E,F) Six1. Arrows, posterior end of Six1expression. (G) RT-PCR of animal caps analysing Six1 andFoxD3 expression. (H,I) Cpl1. (J,K) En2. (L,M) Otx2.(N) Sox2. (O) Keratin. (P,Q) Summary phenotypesproduced by Gbx2 MO (P) and Gbx2 mRNA (Q). Anteriorpart of the embryo is represented, with left side as controland right-hand side as that injected. Percentages ofphenotypes are shown in Fig. S1 in the supplementarymaterial. A minimal of 35 embryos was analysed in eachexperiment. AC, animal cap; E, whole embryo; Gb, 1 ngof Gbx2 mRNA; ODC, loading control; tBR, 2 ng ofdominant-negative of BMP4 receptor.
DEVELO
PMENT
Gbx2 is upstream of the NC genetic cascadeSince the discovery of the transcription factor Slug (Nieto et al.,
1994; Mayor et al., 1995), more than a dozen transcription factors
required for NC development have been identified. Several attempts
to organize these factors into a genetic cascade have been performed
(Mayor et al., 1999; Mayor and Aybar, 2001; Meulemans and
Bronner-Fraser, 2004; Steventon et al., 2005; Sauka-Spengler and
Bronner-Fraser, 2008). From all these proposals the idea emerges
that secreted molecules (BMPs, Wnts and FGF) activate the
expression of a first set of transcription factors, among these Pax3and Msx1. Evidence that any of these factors, or any other known
transcription factor, is directly regulated by the inducing signals has
so far been lacking. Here we show for the first time that Gbx2 is the
earliest element of the cascade, is directly regulated by Wnt signals
and participates in NC induction.
Several lines of evidence indicate that Gbx2 is one of the most
upstream factors in the NC genetic cascade. First, activation of
Gbx2 by Wnt signalling does not require protein synthesis.
Second, Pax3 and Msx1, usually described as the most upstream
factors in the genetic cascade (Sauka-Spengler and Bronner-
Fraser, 2008; Meulemans and Bronner-Fraser, 2004; Steventon et
al., 2005), are downstream of Gbx2. Third, β-catenin/TCF/LEF
factors seem to bind directly to the Gbx2 enhancer. It is interesting
to note that in our ChIP analysis we found that β-
catenin/TCF/LEF does not bind to the Snail2 promoter during
early NC induction, but it binds at later stages. This observation
is consistent with the Snail2 promoter study that shows activity
only after stage 14 (Vallin et al., 2001), and with several
publications that clearly show Snail2 as a factor downstream of
Pax3, Msx1 and Zic1 during early NC induction (Kuo et al., 1998;
Tribulo et al., 2003; Monsoro-Burq et al., 2005; Sato et al., 2005;
Hong and Saint-Jeannet, 2007; Zhao et al., 2008). These data
suggest that inductive signals such as Wnt are likely to work at
different steps during NC development, such as in the early NC
induction by controlling genes such as Gbx2 and later during NC
maintenance by controlling genes such as Snail2. This idea is
consistent with the recent identification of an early induction and
a later maintenance step for NC specification and with the
demonstration that both steps are Wnt dependent (Steventon et al.,
2009).
Gbx2 has been implicated in the formation of the MHB and in the
posteriorization of the neural plate (Joyner et al., 2000). In this work
we show that during gastrulation, when NC induction starts
(Mancilla and Mayor, 1996), Gbx2 is expressed in a broad domain
of the ectoderm including the prospective NC. Later, at neurula
stages, when some additional interactions are required to refine the
position of the induced NC, Gbx2 is absent from the most anterior
NC domain. This loss of Gbx2 is likely to be due to the repressor
activity of Otx2 in the MHB (Glavic et al., 2002). It is known that at
the late gastrula/early neurula stages Gbx2 and Otx2 expression
overlaps in a domain wider than the MHB (Garda et al., 2001;
Glavic et al., 2002). At this stage, Gbx2 is expressed in the entire NC
population, and we show here that it plays an essential role in early
NC induction. Gbx2 knockout mice exhibit defect in NC derivatives,
such as heart and head (Byrd and Meyers, 2005). However, this
phenotype has been explained as a defect in NC pattering or
migration, instead of NC induction. Our results support a role for
Gbx2 in the very early specification of NC cells that explain the
reported deficiencies in NC derivatives in these mutants.
Gbx2 works as a neural fold posteriorizing factorGbx2 has also been implicated in posteriorization of the neural plate,
and its role in hindbrain specification has been widely studied
(Millet et al., 1999; Hidalgo-Sanchez et al., 1999). Our results
support a similar role for Gbx2 as a neural-fold posteriorizing factor.
We show here that anterior neural fold, defined by the expression of
the preplacodal marker Six1, is transformed into NC by the action of
Gbx2, corresponding to the posteriorization process (Nieuwkoop,
3275RESEARCH ARTICLEGbx2 in neural crest induction
Fig. 9. Interaction between Gbx2 and Zic1 is sufficient toinduce NC. (A,B) RT-PCR of animal caps analysing the expressionof the indicated genes at the equivalent of stage 12. (A) Gbx2interacts with a factor induced by attenuation of BMP activity.(B) Interaction of Gbx2 and Zic1 induces NC. Animal capsexpressing the indicated amounts of Gbx2 and Zic1 mRNA.(C-F) Model of NC induction by Gbx2. See text for details. Redasterisks in E indicate the placodes that are at the sameanteroposterior level as the neural crest, such as the otic placode,and are dependent on Gbx2 activity. (F) Network of geneticinteractions that specify PPR and NC. Red arrows, directregulation of Zic1 by BMP (Tropepe et al., 2006) and of Gbx2 byWnt (this work). AC, animal cap; Gbx2, 1 ng of Gbx2 mRNA;Gbx2 MO, 8 ng of Gbx2 MO; ODC, loading control; WE, wholeembryo; Wnt8, 1 ng of Wnt8 mRNA.
DEVELO
PMENT
3276
1952; Cox and Hemmati-Brivanlou, 1995; Villanueva et al., 2002).
Furthermore, we show that Gbx2 is sufficient to repress the
expression of Six1 and promotes the expression of the NC markers
in vitro. Our results provide molecular support to the hypothesis that
NC induction requires posteriorization of the neural fold (Aybar and
Mayor, 2002; Villanueva et al., 2002). Surprisingly we found that
neural fold and neural plate posteriorization can be dissociated:
anteroposterior patterning of the neural fold is possible without
affecting this process in the neural plate. Moreover, the neural fold
seems to be more sensitive to the posteriorizing agents Gbx2 and
Wnt than neural plate (this work) (Carmona-Fontaine et al., 2007).
The different sensitivity offers an explanation to apparently
contradictory results. We have previously shown that addition of
posteriorizing factors, such as Wnt, FGF and RA, transform the
anterior neural fold into NC, supporting a role for posteriorization
in NC induction (Villanueva et al., 2002). This has been challenged
by the observation that activation of Wnt signalling induced NC
markers without any effect in the anteroposterior axis of the neural
plate, concluding that NC induction was independent of
posteriorization (Wu et al., 2005). The results presented here show
that NC induction is independent from neural plate posteriorization,
and that posteriorization of the neural fold can be dissociated from
posteriorization of the neural plate. We propose that NC induction
at the anterior neural fold described by Wu et al. (Wu et al., 2005) is
the result of neural fold posteriorization.
Gbx2 makes the distinction between NC andanterior PPR along the anteroposterior andmediolateral axisThe PPR forms in the outer border of the anterior neural fold and
contributes to sense organs and cranial sensory ganglia (Streit, 2004;
Schlosser, 2006). A recent model for its induction has been
proposed, in which inhibition of Wnt signalling is an essential
component (Brugmann et al., 2004; Litsiou et al., 2005). This model
is consistent with our results showing that Gbx2 is an inhibitor of
PPR and is a Wnt target. Inhibition of Zic1 leads to a depletion of the
NC and PPR population (Hong and Saint-Jeannet, 2007). However,
activation of Zic1 inhibits NC induction and promotes PPR
development (Hong and Saint-Jeannet, 2007). The observation that
activation and inhibition of Zic1 leads to NC inhibition can be
explained with our finding that Zic1 plays a dual role that is context
dependent. In the posterior region of the embryo, Zic1 interacts with
Gbx2 to promote NC specification, whereas anteriorly Zic1 induces
Six1 required for specification of the placode territory (Brugmann et
al., 2004). In addition, different levels of Zic1 may also be important
to specify different territories, as it has been shown for the
distinction between hatching gland, NC and placodes (Hong and
Saint-Jeannet, 2007) and by the direct upregulation of NC markers
(Sato et al., 2005).
We have previously shown that Dkk1 is required to inhibit NC
specification at the anterior neural fold by inhibiting cell
proliferation (Carmona-Fontaine et al., 2007). The data shown here
suggest that in addition to cell proliferation Wnt signalling controls
the specification of NC versus anterior placode. Interestingly, we
have previously shown that placode markers, and not neural plate
markers, are affected by Wnt/Dkk1 (Carmona-Fontaine et al., 2007).
Taken together, these observations suggest that Dkk1 could work as
an inhibitor of Gbx2 at the anterior neural fold region.
We show that injection of Gbx2 leads to a modest lateral
expansion of NC. We propose that this expansion could be due to the
overlap of Gbx2 and Zic1. It is known that Zic1 is expressed in a
wider domain than the NC at the neural plate border that includes
the PPR region (Hong and Saint-Jeannet, 2007), where Gbx2expression is downregulated (Fig. 1T). Thus, injection of Gbx2generates a new region of Gbx2/Zic1 co-expression in the PPR
domain. This new interaction leads to the small mediolateral
expansion of the NC observed. It is interesting to notice that the
same genetic network specifies both the anteroposterior and
mediolateral border between NC and PPR.
The model presented here supports the idea that initial
patterning of the ectoderm is determined by positional information
dependent on two orthogonal gradients. A mediolateral BMP
gradient specifies neural plate, neural plate border and epidermis,
whereas an anteroposterior Wnt gradient divides the neural plate
border into PPR and NC. Interestingly, a similar orthogonal double
gradient of decapentaplegic and Wingless specify the
anteroposterior and mediolateral axis of the Drosophila imaginal
disc (Strigini and Cohen, 1999), suggesting that a BMP/Wnt
orthogonal gradient is an ancient mechanism to generate positional
information.
AcknowledgementsWe would like to thank Eric Theveneau for sections in Fig. 1, Andrea Streit,Claudia Linker, Carlos Carmona-Fontaine and Ben Steventon for comments onthe manuscript, and Laurent Kodjabachian for providing us with the Zic1clone. This study was supported by grants from the MRC and the BBSRC. B.L.was funded by ORS and UCL scholarships. M.M. was supported by an EMBOfellowship. Deposited in PMC for release after 6 months.
Supplementary materialSupplementary material for this article is available athttp://dev.biologists.org/cgi/content/full/136/19/3267/DC1
ReferencesAhrens, K. and Schlosser, G. (2005). Tissues and signals involved in the induction
of placodal Six1 expression in Xenopus laevis. Dev. Biol. 288, 40-59.Aybar, M. J. and Mayor, R. (2002). Early induction of neural crest cells: lessons
learned from frog, fish and chick. Curr. Opin. Genet. Dev. 12, 452-458.Aybar, M. J., Nieto, M. A. and Mayor, R. (2003). Snail precedes slug in the
genetic cascade required for the specification and migration of the Xenopusneural crest. Development 130, 483-494.
Bang, A. G., Papalopulu, N., Kintner, C. and Goulding, M. D. (1997).Expression of Pax-3 is initiated in the early neural plate by posteriorizing signalsproduced by the organizer and by posterior non-axial mesoderm. Development124, 2075-2085.
Basch, M. L., Garcia-Castro, M. I. and Bronner-Fraser, M. (2004). Molecularmechanisms of neural crest induction. Birth Defects Res. C Embryo Today 72,109-123.
Blitz, I. L. and Cho, K. W. (1995). Anterior neurectoderm is progressively inducedduring gastrulation: the role of the Xenopus homeobox gene orthodenticle.Development 121, 993-1004.
Bonstein, L., Elias, S. and Frank, D. (1998). Paraxial-fated mesoderm is requiredfor neural crest induction in Xenopus embryos. Dev. Biol. 193, 156-168.
Bradley, L. C., Snape, A., Bhatt, S. and Wilkinson, D. G. (1993). The structureand expression of the Xenopus Krox-20 gene: conserved and divergent patternsof expression in rhombomeres and neural crest. Mech. Dev. 40, 73-84.
Brugmann, S. A., Pandur, P. D., Kenyon, K. L., Pignoni, F. and Moody, S. A.(2004). Six1 promotes a placodal fate within the lateral neurogenic ectoderm byfunctioning as both a transcriptional activator and repressor. Development 131,5871-5881.
Byrd, N. A. and Meyers, E. N. (2005). Loss of Gbx2 results in neural crest cellpatterning and pharyngeal arch artery defects in the mouse embryo. Dev. Biol.284, 233-245.
Carmona-Fontaine, C., Acuna, G., Ellwanger, K., Niehrs, C. and Mayor, R.(2007). Neural crests are actively precluded from the anterior neural fold by anovel inhibitory mechanism dependent on Dickkopf1 secreted by the prechordalmesoderm. Dev. Biol. 309, 208-221.
Cox, W. G. and Hemmati-Brivanlou, A. (1995). Caudalization of neural fate bytissue recombination and bFGF. Development 121, 4349-4358.
De Calisto, J., Araya, C., Marchant, L., Riaz, C. F. and Mayor, R. (2005).Essential role of non-canonical Wnt signalling in neural crest migration.Development 132, 2587-2597.
Domingos, P. M., Itasaki, N., Jones, C. M., Mercurio, S., Sargent, M. G.,Smith, J. C. and Krumlauf, R. (2001). The Wnt/[beta]-catenin pathwayposteriorizes neural tissue in Xenopus by an indirect mechanism requiring FGFsignalling. Dev. Biol. 239, 148-160.
RESEARCH ARTICLE Development 136 (19)
DEVELO
PMENT
Dunican, D. S., Ruzov, A., Hackett, J. A. and Meehan, R. R. (2008). xDnmt1regulates transcriptional silencing in pre-MBT Xenopus embryos independentlyof its catalytic function. Development 135, 1295-1302.
Eastman, Q. and Grosschedl, R. (1999). Regulation of LEF-1/TCF transcriptionfactors by Wnt and other signals. Curr. Opin. Cell Biol. 11, 233-240.
Gamse, J. and Sive, H. (2000). Vertebrate anteroposterior patterning: theXenopus neurectoderm as a paradigm. BioEssays 22, 976-986.
Garcia-Castro, M. I., Marcelle, C. and Bronner-Fraser, M. (2002). EctodermalWnt function as a neural crest inducer. Science 297, 848-851.
Garda, A. L., Echevarría, D. and Martínez, S. (2001). Neuroepithelial co-expression of Gbx2 and Otx2 precedes Fgf8 expression in the isthmic organizer.Mech. Dev. 101, 111-118.
Ghanbari, H., Seo, H. C., Fjose, A. and Brandli, A. W. (2001). Molecular cloningand embryonic expression of Xenopus Six homeobox genes. Mech. Dev. 101,271-277.
Glavic, A., Gomez-Skarmeta, J. L. and Mayor, R. (2002). The homeoproteinXiro1 is required for midbrain-hindbrain boundary formation. Development 129,1609-1621.
Gomez-Skarmeta, J. L., Glavic, A., de la Calle-Mustienes, E., Modolell, J. andMayor, R. (1998). Xiro, a Xenopus homolog of the Drosophila Iroquois complexgenes, controls development at the neural plate. EMBO J. 17, 181-190.
Graff, J. M., Thies, R. S., Song, J. J., Celeste, A. J. and Melton, D. A. (1994).Studies with a Xenopus BMP receptor suggest that ventral mesoderm-inducingsignals override dorsal signals in vivo. Cell 79, 169-179.
Harland, R. and Weintraub, H. (1985). Translation of mRNA injected intoXenopus oocytes is specifically inhibited by antisense RNA. J. Cell Biol. 101,1094-1099.
Harland, R. M. (1991). In situ hybridization: an improved whole-mount methodfor Xenopus embryos. Methods Cell Biol. 36, 685-695.
Heasman, J., Kofron, M. and Wylie, C. (2000). Beta-catenin signaling activitydissected in the early Xenopus embryo: a novel antisense approach. Dev. Biol.222, 124-134.
Hemmati-Brivanlou, A., de la Torre, J. R., Holt, C. and Harland, R. M. (1991).Cephalic expression and molecular characterization of Xenopus En-2.Development 111, 715-724.
Hidalgo-Sanchez, M., Millet, S., Simeone, A. and Alvarado-Mallart, R. M.(1999). Comparative analysis of Otx2, Gbx2, Pax2, Fgf8 and Wnt1 geneexpressions during the formation of the chick midbrain/hindbrain domain.Mech. Dev. 81, 175-178.
Hong, C. S. and Saint-Jeannet, J. P. (2007). The activity of Pax3 and Zic1regulates three distinct cell fates at the neural plate border. Mol. Biol. Cell 18,2192-2202.
Jonas, E., Sargent, T. D. and Dawid, I. B. (1985). Epidermal keratin geneexpressed in embryos of Xenopus laevis. Proc. Natl. Acad. Sci. USA 82, 5413-5417.
Jonas, E. A., Snape, A. M. and Sargent, T. D. (1989). Transcriptional regulationof a Xenopus embryonic epidermal keratin gene. Development 106, 399-405.
Joyner, A. L., Liu, A. and Millet, S. (2000). Otx2, Gbx2 and Fgf8 interact toposition and maintain a mid-hindbrain organizer. Curr. Opin. Cell Biol. 12, 736-741.
Kawakami, K. (2007). Tol2: a versatile gene transfer vector in vertebrates.Genome Biol. 8 Suppl. 1, S7.
Kishi, M., Mizuseki, K., Sasai, N., Yamazaki, H., Shiota, K., Nakanishi, S. andSasai, Y. (2000). Requirement of Sox2-mediated signaling for differentiation ofearly Xenopus neuroectoderm. Development 127, 791-800.
Kuo, J. S., Patel, M., Gamse, J., Merzdorf, C., Liu, X., Apekin, V. and Sive, H.(1998). Opl: a zinc finger protein that regulates neural determination andpatterning in Xenopus. Development 125, 2867-2882.
La Bonne, C. and Bronner-Fraser, M. (1998). Neural crest induction in Xenopus:evidence for a two-signal model. Development 125, 2403-2414.
Le Douarin, N. M. and Kalcheim, C. (1999). The Neural Crest. Cambridge:Cambridge University Press.
Litsiou, A., Hanson, S. and Streit, A. (2005). A balance of FGF, BMP and WNTsignalling positions the future placode territory in the head. Development 132,4051-4062.
Mancilla, A. and Mayor, R. (1996). Neural crest formation in Xenopus laevis:mechanisms of Xslug induction. Dev. Biol. 177, 580-589.
Marchant, L., Linker, C., Ruiz, P., Guerrero, N. and Mayor, R. (1998). Theinductive properties of mesoderm suggest that the neural crest cells are specifiedby a BMP gradient. Dev. Biol. 198, 319-329.
Mayor, R. and Aybar, M. J. (2001). Induction and development of neural crest inXenopus laevis. Cell Tissue Res. 305, 203-209.
Mayor, R., Morgan, R. and Sargent, M. G. (1995). Induction of the prospectiveneural crest of Xenopus. Development 121, 767-777.
Mayor, R., Young, R. and Vargas, A. (1999). Development of neural crest inXenopus. Curr. Top. Dev. Biol. 43, 85-113.
McGrew, L. L., Takemaru, K., Bates, R. and Moon, R. T. (1999). Directregulation of the Xenopus engrailed-2 promoter by the Wnt signaling pathway,and a molecular screen for Wnt-responsive genes, confirm a role for Wntsignaling during neural patterning in Xenopus. Mech. Dev. 87, 21-32.
Meulemans, D. and Bronner-Fraser, M. (2004). Gene-regulatory interactions inneural crest evolution and development. Dev. Cell 7, 291-299.
Millet, S., Campbell, K., Epstein, D. J., Losos, K., Harris, E. and Joyner, A. L.(1999). A role for Gbx2 in repression of Otx2 and positioning the mid/hindbrainorganizer. Nature 401, 161-164.
Mizuseki, K., Kishi, M., Matsui, M., Nakanishi, S. and Sasai, Y. (1998).Xenopus Zic-related-1 and Sox-2, two factors induced by chordin, have distinctactivities in the initiation of neural induction. Development 125, 579-587.
Monsoro-Burq, A. H., Fletcher, R. B. and Harland, R. M. (2003). Neural crestinduction by paraxial mesoderm in Xenopus embryos requires FGF signals.Development 130, 3111-3124.
Monsoro-Burq, A. H., Wang, E. and Harland, R. (2005). Msx1 and Pax3cooperate to mediate FGF8 and WNT signals during Xenopus neural crestinduction. Dev. Cell 8, 167-178.
Nakata, K., Nagai, T., Aruga, J. and Mikoshiba, K. (1998). Xenopus Zic familyand its role in neural and neural crest development. Mech. Dev. 75, 43-51.
Nguyen, V. H., Schmid, B., Trout, J., Connors, S. A., Ekker, M. and Mullins, M.C. (1998). Ventral and lateral regions of the zebrafish gastrula, including theneural crest progenitors, are established by a bmp2b/swirl pathway of genes.Dev. Biol. 199, 93-110.
Nieto, M. A., Sargent, M. G., Wilkinson, D. G. and Cooke, J. (1994). Control ofcell behavior during vertebrate development by Slug, a zinc finger gene. Science264, 835-839.
Nieuwkoop, P. and Faber, J. (1967). Normal Table of Xenopus laevis (Daudin).Amsterdam: North-Holland Publishing Company.
Nieuwkoop, P. D. (1952). Activation and organization of the central nervoussystem in amphibians. Part III. Synthesis of a new working hypothesis. J. Exp.Zool. 120, 83-108.
Pandur, P. D. and Moody, S. A. (2000). Xenopus Six1 gene is expressed inneurogenic cranial placodes and maintained in the differentiating lateral lines.Mech. Dev. 96, 253-257.
Papalopulu, N. and Kintner, C. (1996). A posteriorising factor, retinoic acid,reveals that anteroposterior patterning controls the timing of neuronaldifferentiation in Xenopus neuroectoderm. Development 122, 3409-3418.
Raven, C. P. and Kloos, J. (1945). Induction by medial and lateral pieces of thearchenteron roof with special reference to the determination of the neural crest.Acta. Neerl. Morphol. 5, 348-362.
Richter, K., Grunz, H. and Dawid, I. B. (1988). Gene expression in the embryonicnervous system of Xenopus laevis. Proc. Natl. Acad. Sci. USA 85, 8086-8090.
Richter, K., Good, P. J. and Dawid, I. B. (1990). A developmentally regulated,nervous system-specific gene in Xenopus encodes a putative RNA-bindingprotein. New Biol. 2, 556-565.
Rohr, K. B., Schulte-Merker, S. and Tautz, D. (1999). Zebrafish zic1 expression inbrain and somites is affected by BMP and hedgehog signalling. Mech. Dev. 85,147-159.
Saint-Jeannet, J. P., He, X., Varmus, H. E. and Dawid, I. B. (1997). Regulationof dorsal fate in the neuraxis by Wnt-1 and Wnt-3a. Proc. Natl. Acad. Sci. USA94, 13713-13718.
Sasai, N., Mizuseki, K. and Sasai, Y. (2001). Requirement of FoxD3-classsignaling for neural crest determination in Xenopus. Development 128, 2525-2536.
Sato, T., Sasai, N. and Sasai, Y. (2005). Neural crest determination by co-activation of Pax3 and Zic1 genes in Xenopus ectoderm. Development 132,2355-2363.
Sauka-Spengler, T. and Bronner-Fraser, M. (2008). A gene regulatory networkorchestrates neural crest formation. Nat. Rev. Mol. Cell. Biol. 9, 557-568.
Schlosser, G. (2006). Induction and specification of cranial placodes. Dev. Biol.294, 303-351.
Selleck, M. A. and Bronner-Fraser, M. (1996). The genesis of avian neural crestcells: a classic embryonic induction. Proc. Natl. Acad. Sci. USA 93, 9352-9357.
Simeone, A. (2000). Positioning the isthmic organizer where Otx2 and Gbx2meet.Trends Genet. 16, 237-240.
Sokol, S. Y. (1996). Analysis of Dishevelled signalling pathways during Xenopusdevelopment. Curr. Biol. 6, 1456-1467.
Steventon, B., Carmona-Fontaine, C. and Mayor, R. (2005). Genetic networkduring neural crest induction: from cell specification to cell survival. Semin. CellDev. Biol. 16, 647-654.
Steventon, B., Araya, C., Linker, C., Kuriyama, S. and Mayor, R. (2009).Differential requirements of BMP and Wnt signalling during gastrulation andneurulation define two steps in neural crest induction. Development 136, 771-779.
Stewart, D., Tomita, A., Shi, Y.-B. and Wong, J. (2006). Chromatinimmunoprecipitation for studying transcriptional regulation in Xenopus oocytesand tadpoles. In Xenopus Protocols: Cell Biology and Signal Transduction,Methods in Molecular Biology Vol. 322 (ed. X. J. Liu), pp. 165-181. Totowa, NJ:Humana Press.
Streit, A. (2004). Early development of the cranial sensory nervous system: from acommon field to individual placodes. Dev. Biol. 276, 1-15.
3277RESEARCH ARTICLEGbx2 in neural crest induction
DEVELO
PMENT
3278
Streit, A. and Stern, C. D. (1999). Establishment and maintenance of the borderof the neural plate in the chick: involvement of FGF and BMP activity. Mech. Dev.82, 51-66.
Strigini, M. and Cohen, S. M. (1999). Formation of morphogen gradients in theDrosophila wing. Semin. Cell Dev. Biol. 10, 335-344.
Suzuki, A., Ueno, N. and Hemmati-Brivanlou, A. (1997). Xenopus msx1mediates epidermal induction and neural inhibition by BMP4. Development 124,3037-3044.
Tada, M. and Smith, J. C. (2000). Xwnt11 is a target of Xenopus Brachyury:regulation of gastrulation movements via Dishevelled, but not through thecanonical Wnt pathway. Development 127, 2227-2238.
Takabatake, Y., Takabatake, T., Sasagawa, S. and Takeshima, K. (2002).Conserved expression control and shared activity between cognate T-box genesTbx2 and Tbx3 in connection with Sonic hedgehog signaling during Xenopuseye development. Dev. Growth Differ. 44, 257-271.
Tribulo, C., Aybar, M. J., Nguyen, V. H., Mullins, M. C. and Mayor, R. (2003).Regulation of Msx genes by a Bmp gradient is essential for neural crestspecification. Development 130, 6441-6452.
Tribulo, C., Aybar, M. J., Sanchez, S. S. and Mayor, R. (2004). A balancebetween the anti-apoptotic activity of Slug and the apoptotic activity of msx1 isrequired for the proper development of the neural crest. Dev. Biol. 275, 325-342.
Tropepe, V., Li, S., Dickinson, A., Gamse, J. T. and Sive, H. L. (2006).Identification of a BMP inhibitor-responsive promoter module required forexpression of the early neural gene zic1. Dev. Biol. 289, 517-529.
Vallin, J., Thuret, R., Giacomello, E., Faraldo, M. M., Thiery, J. P. and Broders,F. (2001). Cloning and characterization of three Xenopus slug promoters revealdirect regulation by Lef/beta-catenin signaling. J. Biol. Chem. 276, 30350-30358.
Villanueva, S., Glavic, A., Ruiz, P. and Mayor, R. (2002). Posteriorization by FGF,Wnt, and retinoic acid is required for neural crest induction. Dev. Biol. 241, 289-301.
von Bubnoff, A., Schmidt, J. E. and Kimelman, D. (1996). The Xenopus laevishomeobox gene Xgbx-2 is an early marker of anteroposterior patterning in theectoderm. Mech. Dev. 54, 149-160.
Wu, J., Yang, J. and Klein, P. S. (2005). Neural crest induction by the canonicalWnt pathway can be dissociated from anterior-posterior neural patterning inXenopus. Dev. Biol. 279, 220-232.
Wurst, W. and Bally-Cuif, L. (2001). Neural plate patterning: upstream anddownstream of the isthmic organizer. Nat. Rev. Neurosci. 2, 99-108.
Zhao, H., Tanegashima, K., Ro, H. and Dawid, I. B. (2008). Lrig3 regulatesneural crest formation in Xenopus by modulating Fgf and Wnt signalingpathways. Development 135, 1283-1293.
RESEARCH ARTICLE Development 136 (19)
DEVELO
PMENT